openoffice-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rob Weir <>
Subject Re: Releasing OOo 3.4 on the old infrastructure
Date Mon, 04 Jul 2011 00:04:50 GMT
On Sun, Jun 19, 2011 at 6:23 AM, Andrea Pescetti
<> wrote:
> I understand that the necessary code reorganization in the Apache
> project will forbid a public release for months.

Aside from the obvious technical tasks of producing code, translation
and documentation, a release at Apache has specific procedural
requirements.  As a Podling project, under incubation, we have
additional requirements.  I'd urge project members to take a quick
read over the following "A Guide to Release Management During
Incubation" for an idea of what is needed:

So yes, a release at Apache will take a good deal of time.  But to me
this is work that is best not delayed.  We cannot avoid it.  We need
to deal with these topics, and the earlier we start, the sooner we
will finish.

> On the other hand, 3.4 beta has been out for several
> weeks, is regarded by our QA testers as remarkably stable, has a
> dedicated code line and it is not far from release.

This is great to know.  This means that we will not need to spend a
lot of time making the code stable.  But we'll have other things to
do, like removing GPL dependencies, etc.  So some I'd expect that
changing of that magnitude will introduce (at least initially) new
bugs.  So it might be reasonable to have a 2nd beta once we have a
clean build.

> Would it be possible to release OOo 3.4 on the old (Oracle-owned)
> infrastructure, and maybe take advantage of this release to educate
> users and volunteers about the coming new infrastructure at Apache?

I don't see how this could be possible for this Apache project to do this.

> This shouldn't require an exceptional effort and it would allow to show
> that it in good shape and to use the momentum of a new
> release to communicate better the transition to Apache. Otherwise we
> really risk to confuse users and let them feel abandoned, and work on
> the incubator in the next months could be less useful.

If user confusion is a concern, maybe we should address that directly?
 For example, it is possible for this project to have a blog, where we
could explain what were are doing and what the plan was.  This
assumes, of course, that we first agree upon a plan.

> It really seems a small effort compared to the major code refactoring
> needed at Apache, and I believe Oracle can state that the bugfixes on
> the OOo 3.4 code line will be granted to Apache too. And the benefits
> for users and communication would be immense.

I don't disagree with you that completing OOo 3.4 on the existing
infrastructure would be easier.  I just believe that we should not
avoid the hard work of migration any longer.

> Now, I take for granted that the community would support this proposal
> (for one, the Italian community spent weeks to get the OOo 3.4 strings
> 100% translated into Italian, and our QA team is ready to start full
> testing any moment). Would developers and release managers support this
> too?

Again, I don't see how it would be possible for an Apache project to
do this.  Of course, if a group outside of Apache wishes to do
something, with the current infrastructure and current code with
current license, then that is a different question.  But maybe you
should bring this up on the OOo mailing list?



> Regards,
>  Andrea.

View raw message